How We Tamed PII in AI CareerBot
In AI CareerBot, users often send resumes containing full names, phone numbers, and addresses. We implemented a PII handling policy in line with Federal Law 152-FZ: we keep the conversation context, but before sending anything to the GigaChat API, we filter out personal data.
In AI, everyone is rushing to ship chatbots, but there’s a boring, yet critical, adult problem: personal data. In AI CareerBot, users often paste chunks of resumes with full names, phone numbers, and addresses — the usual story: “evaluate my experience,” and suddenly half their life is in one message 😅
We implemented a PII handling policy in line with Federal Law 152-FZ: we keep the conversation context so that follow-up messages don’t turn into “who are you and what were we talking about?”, but before sending anything to the GigaChat API, we filter out personal data. In other words, the meaning, structure, and usefulness of the conversation stay intact, while sensitive data does not leave the system.
In practice, we struck a balance between AI convenience and privacy by design — because in 2025, that’s no longer a nice-to-have, it’s the baseline. Especially when regulation tightening is not news, but everyday reality.
The moral: a good AI conversation is not just smart answers, but also proper data hygiene.